Proposed is a program of training and research using ongoing longitudinal studies to investigate the effects of dynamic environments on childhood obesity. Obesity and sedentary behavior are increasingly common threats to childhood health, likely reflecting changes to the physical activity opportunities, food environment, and social contexts experienced by children and their families. Characteristics of the built environment have been linked to health behaviors and weight of both children and adults, but the causal nature of the link is not clear. A key challenge is to distinguish between competing explanations for the observed associations: the effect of neighborhoods on individual physical activity and health versus the tendency for active and healthy individuals to select neighborhoods that support their lifestyle. The proposed research has the potential to corroborate or temper the conclusions of previous research on this topic. Whereas previous cross-sectional studies have largely assumed that residential environments are static, the proposed project will use data on residential mobility and dynamic environments to better establish the temporal sequence of purported environmental causes and health effects. Available social context measures will be leveraged to show how built environment associations compare to or interact with those of social context. I will select analytical approaches to take full advantage of these data and to assess competing causal hypotheses. I will analyze data on built environments, social context, physical activity, and weight from two longitudinal studies. The first is the Fragile Families Study of Childhood Wellbeing, a birth cohort study of 4,898 children in large US cities that oversampled single parent families and is currently collecting data at age 9. The second is the Healthy PLACES trial a natural experimental evaluation of a newly built Smart Growth development, with built environment characteristics expected to have beneficial effects on local social networks and physical activity from parent-child pairs. Analyses in these cohorts will address specific hypotheses focusing on the obesogenic effects of the built environment, social context, and their interaction. In addition to the use of longitudinal and natural experimental designs, I will explore a range of analytical strategies selected to test the causal nature of observed associations. In particular, I will examine sequencing, specificity, and selection through stratified analyses, structural equation modeling, latent trajectory analysis, directed acyclic graphs, propensity score matching, and instrumental variable analysis. The proposed research will be complemented by experiential and didactic training in (1) environmental approaches to decrease health disparities, (2) theory and measurement of social context and health, (3) trajectory analysis and lifecourse epidemiology, and (4) study designs and analytical approaches to inform causal inference. This research and training will enable me to launch a program of independent research on the urban built and social environments shaping childhood physical activity and health.